Combining Belief Function Theory and Stochastic Model Predictive Control for Multi-Modal Uncertainty in Autonomous Driving

TL;DR

The paper tackles multi-modal uncertainty in autonomous driving by integrating Belief Function Theory (BFT) with Stochastic Model Predictive Control (SMPC) to design collision-avoidance constraints that reflect estimation reliability. It introduces two mechanisms: an inverse plausibility transformation that converts BFT beliefs into trajectory probabilities used as SMPC risk parameters, and a constraint tightening scheme that scales the safety constraint based on the BFT reliability $\mu$ and plausibility $\mathrm{Pl}(\{\theta_i\})$ with a tunable $0<\gamma<1$. The methods are evaluated in highway and urban intersection simulations, showing that accounting for estimation reliability improves safety without unnecessary conservatism when intentions are clear, and increases caution when uncertainty is high. The results demonstrate that BFT-informed SMPC can balance safety and efficiency in unstructured traffic environments, providing robust decisions amid ambiguous TP behavior. Overall, the proposed framework offers a practical path to more reliable AV trajectory planning under epistemic uncertainty.

Abstract

In automated driving, predicting and accommodating the uncertain future motion of other traffic participants is challenging, especially in unstructured environments in which the high-level intention of traffic participants is difficult to predict. Several possible uncertain future behaviors of traffic participants must be considered, resulting in multi-modal uncertainty. We propose a novel combination of Belief Function Theory and Stochastic Model Predictive Control for trajectory planning of the autonomous vehicle in presence of significant uncertainty about the intention estimation of traffic participants. A misjudgment of the intention of traffic participants may result in dangerous situations. At the same time, excessive conservatism must be avoided. Therefore, the measure of reliability of the estimation provided by Belief Function Theory is used in the design of collision-avoidance safety constraints, in particular to increase safety when the intention of traffic participants is not clear. We discuss two methods to leverage on Belief Function Theory: we introduce a novel belief-to-probability transformation designed not to underestimate unlikely events if the information is uncertain, and a constraint tightening mechanism using the reliability of the estimation. We evaluate our proposal through simulations comparing to state-of-the-art approaches.

Figures (8)

• Figure 1: Initial traffic configuration and candidate trajectories for the highway scenario.
• Figure 2: Belief assignment for the candidate TP trajectories.
• Figure 3: Trajectories of TP (dashed) and of the EV (solid) resulting from different collision-avoidance constraints.
• Figure 4: Initial traffic configuration and candidate trajectories for the urban intersection scenario. TP1 and TP2 could turn right, continue straight or take the bike lane on the other side, TP3 could cross vertically or horizontally.
• Figure 5: Belief assignment for the candidate TP trajectories.

Theorems & Definitions (2)

• Theorem 1
• proof